Show HN: We Beat MLPerf: Modern Storage for KV Offload and LLM Training

MLCommons released MLPerf Storage v3.0 results on September 1, 2026, an industry-standard benchmark for storage performance on AI workloads. OpenLake achieved the highest write and read bandwidth among published Closed division S3 results for the Llama 3.1 8B checkpointing workload using one client node and one data parallel instance. The announcement highlights that these five submissions are the closest like-for-like comparison available, though underlying storage configurations differ.